Is BPM the killer app for agents No
Is BPM the killer app for agents? No Why not? • Business reasons: BPM is already being done by IBM, Microsoft, BEA, Oracle, and a host of other small/mid/large enterprise SW vendors • Agents provide little new end-functionality – at most “minor” architectural improvements But, there’s a potential niche opportunity for agents in BPM • Off-line participation in business processes with seamless reconnectivity to BPM server when on-line • Better: ad-hoc BPM where servers infeasible (construction sites, tractors) © 2004, Quantum Leap Innovations. All rights reserved.
Issues BPM Challenges & pitfalls • Current approaches have highly static architecture & interfaces (despite promise of web services) Agents bring • (almost) zero configuration enables highly dynamic (possibly even adaptive) architectures • Low cost scalability • Centralized control not a fundamental problem However, the agent improvements are not sufficient to counter the established vendors (from business point of view) • Established markets • Established (centralized) architectures © 2004, Quantum Leap Innovations. All rights reserved.
The Killer App is… I don’t know But characteristics will be: • Something ONLY an agent-oriented approach can achieve • Complex, rapidly changing environment • Driven by “business” – not technology In the meantime • BPM definitely provides good application & research area for agents © 2004, Quantum Leap Innovations. All rights reserved.
- Slides: 3